JPH06205027A - Local loop back of isochronal data in switching mechanism - Google Patents

Abstract

PURPOSE: To keep a large band width margin on the bus of a large band width for other isochronous or non-isochronous data traffic. CONSTITUTION: A data communication system such as a LAN or a WAN is provided with isochronous data transmission ability. The system time-division multiplexes both isochronous and non-isochronous data to a repetitive frame structure with a 4-bit nibble as a base and transmits them. Arrived data are demultiplexed to different channels, so as to process individual data streams by an appropriate hardware in a hub, for instance. The data are hierarchically passed from sources 48c and 48d via a node 42b to the hub. The hub places the data in an internal connection memory, switches them onto the bus of the large band width and distributes them to the other destination station hub, the node or the combination of sink. In the case that a source node and a destination station node are attached to the same hub, the hub directly performs local loop-back to the destination station node, and the need for placing transmission data on the bus first is avoided.

Description

Detailed Description of the Invention

[0001]

FIELD OF THE INVENTION The present invention relates to data communication networks such as local area networks or wide area networks, and more particularly to networks for transferring isochronous data.

[0002]

BACKGROUND OF THE INVENTION Isochronous data can generally be described as unpacketized, indeterminate, potentially continuous length data. Isochronous data sources include video cameras that output a substantially continuous stream of data representing images and associated sounds, and telephones that output a substantially continuous stream of audio data. An example of an isochronous data sink is a video monitor that receives a substantially continuous stream of video data for display.

FIG. 1A schematically illustrates isochronous data transfer. The transfer or "connection" of data is first initiated, for example, by initiating a telephone conversation or by initiating a video camera transfer 12. Probably housekeeping information (eg destination station, source, audio or video timing etc.) after the data transfer is initiated.
The transfer of data, which is accompanied by the transfer of, occurs substantially continuously over an indeterminate period, for example up to the end 14 of the telephone conversation or video transmission. Not all bits transferred need represent data bits. "Housekeeping" bits that control the destination station and audio or video timing may also be transferred. Furthermore, the data transferred can consist of "empty" data, such as silence during a telephone conversation, or transfer of a blank video image. One form of isochronous data transfer is
For example, FDDI-II Hybrid Multiplexer on March 25, 1991,
Fiber Distributed Data Interface-II (FDDI-II), as described in Revision 2.4.

Due to the increased availability of multimedia computers and workstations containing isochronous sources and sinks in addition to non-isochronous sources and sinks, there has been an interest in transmitting isochronous data in a network environment. It is rising. Many existing networks use non-isochronous data communication between stations on the network. Commonly used data transfer protocols include packet transfer systems and token ring systems.

One example of packetized data transfer is the commonly used Ethernet system. 10BASE
One embodiment known as -T is November 15, 1989.
It is described in Draft 9, supplemented by IEEE Standard 802.3 of the day. FIG. 1B shows the packet transmission 22.

In token ring systems, nodes transfer data only after capturing an electronic token. One commonly used token ring system is described in IEEE Standard 802.5. Figure 1
(C) is data transfer in the token ring system 23
Is shown.

[0007]

Many prior attempts to adapt isochronous data over these existing data networks have resulted in adverse operating characteristics. In some conventional isochronous devices, the available bandwidth for a given isochronous source or sink is directly related to the total number of isochronous sources and sinks transmitting and receiving on the network. Decrease proportionally. Furthermore, the presence of isochronous sources and sinks reduces non-isochronous bandwidth.

In addition, existing isochronous systems also provide little or no compatibility with conventional networks. This incompatibility requires extensive exchange of hardware or software to accommodate both isochronous and non-isochronous traffic. Thus, a multimedia personal computer with Ethernet capability and a video camera cannot simultaneously use both isochronous and non-isochronous sources / sinks.

[0009]

Applicant's U.S. patent application Ser. No. 969916 entitled "Data Communication Network With Isochronous Capabilities" filed on the same day and incorporated herein by reference. Issue describes a system that provides data communication to and from isochronous data sources and sinks. The bandwidth available for isochronous sources / sinks is independent of changes in non-isochronous demand on the network and vice versa. Furthermore, each source / sink is guaranteed isochronous bandwidth, which is independent of changes in the source / sink bandwidth on the network. Isochronous communication systems also maintain a high degree of compatibility with conventional, often installed systems, requiring minimal hardware / software exchanges. For example, the effective data rate of isochronous data is a change in bandwidth or traffic of non-isochronous data, or interruption in non-isochronous data (data collision in Ethernet data or loss of token in the case of token ring data). Depending on, it does not change.

Preferably, the system of the present invention is implemented as a star network, where the data source forwards to a central hub, which in turn forwards the data to a data sink. A hub can be a high bandwidth bus, such as a timeslot interchange (TS
I) It is possible to connect several such star systems by interconnecting to the bus, for example in a ring or tree structure. The multiplexed data that arrives at the hub is demultiplexed to separate isochronous source data, non-isochronous source data, and D and M channel information. The data from this non-isochronous source can be provided to a hub circuit that is specialized to handle the non-isochronous data stream. Preferably, circuitry in the hub converts the separated non-isochronous data stream into a format substantially similar to that available on conventional non-isochronous networks. For example, if the non-isochronous data is Ethernet M
If derived from AC, the hub converts the separated non-isochronous data into a form that is processed by a standard Ethernet hub repeater circuit.

Similarly, isochronous source data can be provided to a hub circuit that is specialized to handle isochronous data streams. According to one embodiment of the present invention, the hub isochronous data circuit has local loopback capability. In the situation where the source node and the destination station node are coupled to the same hub circuit, local loopback performance allows data to be transferred between nodes without the hub first putting the data on the TSI bus interconnecting the hubs. To be able to transfer in. As a result, the bandwidth of the TSI bus is not affected by the internal transfer. Thus, the system's ability to interconnect nodes on different hubs is not affected by transfers between nodes connected to the same hub.

According to another embodiment of the present invention, a local loopback has a source node and a destination station node connected to the same hub, where the hubs are connected to each other via a high bandwidth bus. It can also be used in various network configurations. Thus, the present invention provides a system for completing such transfers without affecting the bus external bandwidth.

[0013]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT Table 1 at the end of this specification is a table of time division multiplexing schemes for multiplexing data streams according to one embodiment of the present invention. Table 2 also lists the forms of 4/5 encoding according to one embodiment of the invention.

A detailed description of the present invention is provided with respect to a data transfer system that supports both non-isochronous and isochronous communication. The description given here is therefore 1)
This book describes transfers between isochronous sources and isochronous sinks connected to the same hub, and 2) transfers between isochronous sources and isochronous sinks connected to the same hub. It describes the case where the invention is used. The description herein will therefore show some of the more general situations in which the invention may be used. However, the present invention is not limited to use only in the particular isochronous / non-isochronous data transfer system described, but for any data transfer system in which communication between nodes is via a hub. Can be easily adopted for.

Applicant's US Patent Application No. 969916, entitled "Data Communications Network with Isochronous Capability," filed on the same day and incorporated herein by reference, is in star form. It describes a data communication system for isochronous data that can be arranged and interconnected in a ring or tree configuration. Such a system is shown in FIG. 2, FIG. 3 or FIG. In the arrangement shown in FIG. 2, the hubs are connected in a ring, with the first hub 44a sending data to the second hub 44b and the second hub 44b to the third hub 44c. And the third hub sends the data back to the first hub 44a via the cycle generator and the ring latency adjustment circuit. The connections between the hubs are made on a timeslot interchange (TSI) ring 58f. In one embodiment, the TSI ring 58f is F
A DDI-II system can be used. FIG. 3 shows hubs 44a, 44b and 44c arranged in a star and ring configuration with multiple isochronous circuits within a single hub. FIG. 3 shows a tree-type communication system. Parent hub 44
a connects to a high bandwidth backbone. Hub 4
4b acts as a child hub of parent hub 44a and is attached to port 2 of hub 44a. The child hub 44c is cascaded from the child hub 44b.

The star and ring configurations are multiple nodes 42a, 4 attached to a single hub operating on a high bandwidth bus.
Includes 2b and 42c. The exact number of nodes will vary depending on the need for data transmission and the purpose of the system. Each of the nodes 42a-42c has various types of sources and sinks, such as strict isochronous sources and sinks, strict non-isochronous sources / sinks, or both isochronous and non-isochronous sources and sinks. Can be included. A data link consisting of a physical data transmission medium, such as a one-way twisted pair cable 46a-46r, connects each node to a hub 44a-44.
is bound to one of c.

FIG. 5 illustrates hub 44a and associated nodes 42a-4.
2c is shown in more detail. FIG. 5 can itself form a complete star system. Each node 42a, 42b,
42c includes circuits 50a, 50b, 50c. Circuit 50a-c
Receives the data and transforms it into a form suitable for transmission on physical media 46a, 46c, 46e, and physical media 46b, 46d, 46e.
Receives a signal from f and transforms it into a form suitable for use by a data sink.

Hub 44a receives data from physical media 46a, 46c, 46e and converts isochronous data to non-isochronous data and D
It includes circuits 54a, 54b, 54c for separating the channel and M channel data and converting the separated data into a form suitable for processing by the downstream hub circuit 56. In the illustrated embodiment, the data from the isolated isochronous source is fed to an isochronous switching circuit, such as timeslot interchange controller 58, to place the data on the TSI bus, thus Destination station node
It can be transmitted to and recovered from other hubs by other equivalent circuits 54a-54c at that hub for transmission to 42a, 42b, 42c. The separated non-isochronous data is
It is provided to a circuit 60 configured to carry non-isochronous data for transmission to destination station nodes 42a, 42b, 42c. In embodiments where the data from the non-isochronous source comprises Ethernet data, hub circuit 60 may be a standard Ethernet repeater processor. In this way, the system of the present invention may be at least partially backward compatible with conventional Ethernet hub systems.

The D channel and maintenance data are provided to the signaling processor 62. Signaling processor 62 performs various maintenance and control functions. For example, identify error conditions and alert the user, for example, data path 64
It communicates with the isochronous and non-isochronous controllers 58, 60 above to set up the requested connection, ie the source / destination station path.

The operations of the above-described components are the data transfer from the video camera 48d which is an isochronous source to the isochronous sink 48b, and the Ethernet MAC which is an isochronous source.
It will be understood by describing the data transfer from 48c to non-isochronous sink 48g. The data emitted from isochronous device 48d is a continuous stream of digitized data, having a data rate equal to, for example, the American "T1" standard of 1.544 Mbps. The data output from the Ethernet MAC 48c is provided at a standard 10BASE-T Ethernet transfer rate of 10 Mb / sec. The D channel information is
It is preferably provided from a D channel data stream source included in circuitry such as a MAC or other system, or from a virtual keypad 48f at a variable data rate, such as a rate not exceeding about 64 Kbps.

Lines 66a, 66b, 66c include sources 48d and 48.
The data stream from c is carried to node circuit 50b. FIG. 6 shows this circuit 50b in more detail. The node circuit 50b includes hardware that operates on the incoming data stream to enable efficient, compatible transmission between the data source and the destination station. The multiplexer 70 time-division multiplexes the input data in units of 4 bits by using a series of frames or repetition of templates. In this example, the frame is
Repeated every 125 microseconds.

Table 1 shows the various data streams and the manner in which the additional data and control bytes are time division multiplexed. Each symbol in Table 1 represents 4-bit data, and one data byte of 8 bits is represented for each group of two symbols. In Table 1, E represents 4-bit data from Ethernet stream 66a, B represents 4-bit data from isochronous stream 66b, and D represents 4-bit data from signaling or D-channel stream 66c. M preferably represents the 4-bit M channel data provided by circuit 50b. In addition, a pattern of some byte length is introduced. JK represents a frame synchronization pattern, and EM
The (first two bytes of block 3 in Table 1) represent the Ethernet "pad", followed by the maintenance byte.

As seen in Table 1, each frame is
It contains 256 bytes, which can be thought of as 32 groups of 8 bytes each, or 4 groups of 64 bytes each. For output at a data rate of 1.544 Mb / sec from isochronous source 48d, the frame structure described above yields an isochronous bandwidth capability of 6.144 Mb / sec. Thus, the single isochronous source 48d in this embodiment can be fully accommodated using only 192 "B" symbols per frame. A basic rate ISDN channel can be supported by using three 64 Kb / s in the isochronous channel. Thus, various isochronous sources can be allocated within the available isochronous bandwidth. This frame structure was filed on the same date and incorporated herein by reference, the contents of which are incorporated herein by reference, the applicant's U.S. patent application Ser. Is described more completely in. Other frame structures different from the ones described above can be used to provide a bandwidth allocation suitable for a particular purpose.

The time division multiplexed data is then encoded by the encoder.
The AC balance of the cable is maintained, which is encoded by 72 and can be confused by long strings of binary zeros. In the illustrated embodiment, the encoder performs 4/5 encoding. One particular form of 4/5 encoding, which partially conforms to the ANSII X3T9.5 standard, is shown in Table 2.
Is shown in. When properly combined, these patterns have a maximum of 3 bit times and no transitions. The encoding scheme shown in Table 2 is more commonly found in Applicant's U.S. Patent Application No. 970329, entitled "Frame-Based Data Transmission", which was filed on the same date and incorporated herein by reference. Described in detail.

The result of the 4/5 encoding is then Non-Re
It is further encoded by the encoder 74 of FIG. 6 using the turn to Zero Inverted (NRZI) method. This 4,5-NRZI encode is non-isochronous source 10BA
In the network that is the SE-T Ethernet source,
Especially useful. The reason is that this encoding provides transmission at signaling rates that are substantially compatible with the expected data rates provided by Ethernet MACs. However, other forms of encoding or decoding can also be used, such as encoding 8 bits to 10 bits.

After encoding, the data is sent to the pre-emphasis circuit 76 and the transmitter or driver 78b. The pre-emphasis circuit 76 compensates the signal transmitted on the physical medium to reduce the jitter. The signal is then
It is transmitted to the hub 44a via the physical medium 46c, and the physical medium 46c includes a twisted pair cable, a coaxial cable,
Alternatively, an optical fiber cable or the like is included.

The hub 44a seen in FIG. 5 includes a plurality of circuit devices 54a, 54b, 54c, each of which is a physical medium 4.
It is coupled by 6 to one of nodes 42a, 42b, 42c. As shown in FIG. 7, data transmitted over physical medium 46 arrives serially at deserializer / decoder 80. The deserialization circuit / decoder 80 includes a circuit that is functionally the reverse of the above described multiplexing / encoding circuit to decode the 4/5 NRZI encode to separate the isochronous and non-isochronous source data. To work. Deserializer / decoder 80 also outputs a sync signal derived from JK frame sync symbol 96 for use by framing timing generator 98. Link detection circuit
82 also receives data from the physical medium 46 to detect the mode in which the node is operating (eg, 10BASE-T, non-isochronous Ethernet or isochronous), filed on the same date and referenced herein. A mode select signal is output as described more fully by Applicant's U.S. Patent Application No. 971018 entitled "Network Link Endpoint Capability Detection", the contents of which are incorporated herein.

Both the non-isochronous source data 94b and the isochronous source data 94a are optionally sent to various hub circuit components 54a, 54b, 54c for transmission to the destination station node.
Is made available to. In one embodiment,
The separated isochronous data 94a and non-isochronous data 94b are
Reconfigured by interfaces 58 and 60 respectively,
Isochronous output 102 and non-isochronous output 104 are provided in a form suitable for transmission to a destination station node. In one embodiment, the non-isochronous data 94b is configured by the E-interface 60 to be included in the hub circuit 54 so that the output data 104 can be processed by the repeater device for final transmission to the destination station node. To be. As an alternative to using repeaters for non-isochronous data, packet connections can be linked via media access control layer bridges.

FIG. 8 illustrates an E interface 59 in the form of receiving non-isochronous data 94b and providing outputs 106, 108 in a form that can be processed by previously available repeater circuit 60.
2 illustrates one embodiment of This non-isochronous data is received in a first in first out (FIFO) buffer 112 to smooth the data transfer rate. Circuit 114 is
It detects the "no carrier" symbol provided to emulate an Ethernet data packet, which is used by a logic circuit or state machine 116 to output a carrier detect signal. The output 118 from the FIFO 112 is provided to a multiplexer 120 and a deserialization circuit 122,
A data output 106 is produced. Multiplexer 120 can receive preamble stream 124 and provides the appropriate preamble bits in output data 106. F
The output 118 from the IFO 112 is also provided to a decode circuit 128 to recognize data collision and alignment error symbols and output appropriate signals 130, 132 to the state machine 116. The operation and components of receive interface 59 are more fully described in US Patent Application No. 970329, entitled "Frame-Based Data Transmission."

For the purposes of this example, an isochronous source
The data from 48d (FIG. 9) is represented by the "B" symbol in block 0 of Table 1 of the first 24 isochronous bytes of each frame (ie the first 48 "B" in the frame structure).
Symbol) is assumed to be transmitted. FIG. 9 illustrates a B interface 58 according to one embodiment of the present invention. In the example of FIG. 9, separated isochronous data
94a is stored in one of the two buffers 132a and 132b. The timing of storage in buffers 132a, 132b has been adjusted with a frame transmission timing of 125 microseconds so that the data 94a from the first frame is stored in the first buffer 132a during the first 125 microsecond period. , The isochronous data 94a from the next frame is stored in the second buffer 132b during the next 125 microsecond period. In one embodiment,
The data may be stored in buffer 132 in the same order as it was received, with the 8 bits represented by the first two "B" symbols in Table 1 being stored in the first storage location of buffer 132a, Those corresponding to the next two "B" symbols in Table 1 are stored in the second location of buffer 132a, and so on. Since the frame structure shown in Table 1 contains 96 bytes of isochronous data per frame, each of the buffers 132a, 132b has the ability to store 96 bytes of data per node it supports. . Isochronous data from the first frame is buffered 132
After being stored in the first buffer 132a during the next 125 microsecond period (while the data from the next frame is being stored in the second buffer 132b), the data stored in the first buffer 132a is It is transmitted on high bandwidth bus 134.
The loading and ordering of buffer 132 depends on the number of nodes supported by hub 44a. Bus 134 has sufficient bandwidth to carry the isochronous data outputs from the nodes connected to hub 44a. In the embodiment where hub 44a is connected to 16 nodes, the bandwidth of bus 134 receives 1536 bytes of data every 125 microseconds (ie, per frame) (ie, 96 bytes per node × 16 nodes). Must be enough to do. This corresponds to a bandwidth of about 98304 Kb / s.

Depending on the aspect of the system configuration, such as the number of nodes attached to the hub and the bandwidth dedicated to isochronous data, other embodiments of the present invention may have other bandwidth alternatives to the TSI bus 134. Can be provided. However, the bandwidth of 98304 Kb / s is particularly useful. Because it substantially matches the bandwidth used in FDDI-II, and in a configuration where TSI ring 58 is an FDDI-II system, the data on TSI bus 134 is
This makes it particularly easy to transfer to the ring 58 (FIG. 5).

In one embodiment, data is carried from buffer 132 into time slots on bus 134 in a time slot interchange fashion. TSI bus 134
The data carried above is transmitted in a 125 microsecond time frame divided into 1536 time slots,
Each of them has a length of about 0.08138 microseconds. Each timeslot has data and associated control and parity. Thus, one byte can represent 10 bits of time slot information. Thus buffer 13
Data from 2a goes onto TSI bus 134 and into buffer 132a
Is placed on the TSI bus 134 in the appropriate slot of the 1536 time slots of the 125 microsecond time frame. Which time slot is the “right”
It depends on the application for which the data is used, and in particular on the predetermined destination station of the data in the connection setup via the D channel.

In the illustrated embodiment, the data destination station is pre-established using the D channel information. This D channel information is sent to the signaling processor 138. This D-channel information, including source, destination station and other necessary information, is preferably a switch table.
Used to store the value in 140. In one example,
Switch table 140 is divided into 16 portions 142a-142p corresponding to the 16 nodes associated with hub circuit 58 in this example. Each portion 142 contains 1536 bits corresponding to 1536 time slots in the time frame of the TSI bus. These bits are
Can be used as control 144 for.

In this embodiment, the 24 bytes of data from isochronous source 48d per 125 microsecond frame is the first 24 frames of each isochronous source 48d frame.
In the B slot. Therefore, the source 48d
Data is stored in the isochronous buffer 132. The destination station for isochronous data in this example is monitor 48b. Therefore, the 24 B slots of data are the data buffer 15
4a and then on the next frame to sink 48b in its corresponding first 24 B slots.

The 24 B slots can be directed to the TSI bus, in which case the 24 B slots of the isochronous buffer 132 are switched onto the TSI bus. A bit of the contents of the switch table controls line 150 and controls multiplexer 146 at a rate of 1 bit per TSI timeslot (ie 1 bit every 0.08138 microseconds).
The first 10 timeslots of the TSI bus are the first TSI
Assuming that it does not receive B data destined for a node attached to another hub during the timeslot, multiplexer control 114 is a "0" and no data is output from buffer 132 to bus 134. Multiplexer 146
Merely conveys along the TSI bus 134 some data already on the TSI bus in the first time slot. This condition continues until the 11th time slot on the TSI bus, at which point B data destined for the node attached to another hub begins to be output on the TSI bus. On each of the next 24 TSI bus time slots, the control signal to multiplexer 146 is a "1" and the data byte stored in the appropriate data location of buffer 132 is from multiplexer 146 to bus 1.
34 is output up. Which data position in the buffer 132 is "suitable" can be determined by the read pointer contained in the switch table. Preferably, the buffer 132 is a random access memory (RA
M) and the read pointer is determined according to the contents of the switch table position representing the TSI slot frame. After the transmission on the 24-byte TSI bus is complete, there is no output from buffer 132a during the subsequent timeslots of this TSI frame, because no other connection has been established in this example. . .
In this way, time slots 11-35 for a frame on the TSI bus are filled with the data stored in buffer 132a, i.e., the 24-byte data output by isochronous source 48d.

FIG. 9 also illustrates the transfer of isochronous data recovered from the TSI bus 134 to the destination station node. In this embodiment, it is necessary for hub 44a to recover the 24 bytes of data stored in the first 24 even time slots of the transmitted frame. The data from the TSI ring is the B interface 5 associated with sink 48b.
Recovered by 8.

Recovery from the TSI ring depends on the table 162 in a similar manner as described for the control of the multiplexer 146 from the signal processor 138 to the line.
This is accomplished by multiplexer 156 controlled by control signal 158 output via 160.

The E interface 60 of hub 44a recovers the non-isochronous data (source 48c) from repeater 60 intended for non-isochronous sink 48g. E transmission interface 168
An example of is shown in FIG. The transmission interface shown in FIG. 10 is generally functionally the reverse of the E reception interface shown in FIG. It is also possible to provide a parallel interface and there is no need for a FIFO when in the MAC. Data 166 is deserialized, then any required alignment error bits 172 and multiplexer 1
Combined at 74, the output is output to the FIFO 176. The sync detection circuit 178 extracts the sync information from the repeater output 166 and conveys it to the state machine 180. State machine 180 also receives carrier detect information 184, framing counter information 186, and sends control signal 188 to FIFO 176.
To provide. The data output from the FIFO 176 is multiplexed with the preamble bit 190 and the "0 carrier" bit 194 by the multiplexer 196. The operation of the E-Transmission interface is more fully described in US Patent Application No. 970329, entitled "Frame-Based Data Transmission."

The data 198 output from the E transmission interface 168, along with the isochronous data output 164 and the M and D channel data 170, is provided to the encoder serialization circuit 202 as shown in FIG. The encoder / serialization circuit 202 is constructed in substantially the same manner as the encoding circuit shown in FIG. Specifically, the encoder / serialization circuit 202 includes a multiplexer for combining three streams of data 198, 170, 164,
It provides a 4/5 encoder, an NRZI encoder and a pre-emphasis circuit. The transmission timing is controlled by the transmission timing circuit 204. The output 206 from the encoder / serialization circuit is a link beat for the purpose of link endpoint detection by a multiplexer 210, as more fully described in Applicant's US patent application No. 971018. Selectively coupled to the link beat from generator 208.

Both isochronous and non-isochronous data sent from the hub 44a to the node 42 is stored in the node as described above.
It is sent in a frame format that is preferably substantially the same as the frame format used for the data sent from 48 to hub 44a. Node 42
In FIG. 6, the circuit 50 uses a device similar to the device described above for performing these functions in the hub for decoding and demultiplexing data (FIG. 6), mainly a phase-locked decode circuit 86, an NRZI decode circuit 88, 4. / 5
It includes a decoding circuit 90 and a demultiplexer 92. The decoded and demultiplexed data is then communicated to various data sinks at node 42.

FIG. 12 shows a timing scheme for reducing attenuation, jitter, and minimizing the amount of buffering memory required. As shown in FIG. 12, this timing can be synchronized with the 125 microsecond reference clock signal 214, which results in a rising clock edge every 125 microseconds. This reference signal can be provided by any of a number of sources, including synchronization with an external clock reference, such as a reference signal from a wide area network or FDDI-II ring. At the beginning of the cycle, hub 44 begins transmitting frames to the node, as indicated by the timing marks on timeline 216. Due to the line delay in the physical medium, as indicated by time line 218, the time points at which the nodes receive the frames transmitted by the hub are delayed from the time they are sent from the hub.
There, a delay 220 is introduced before the node starts transmitting the next frame to the hub 222. This delay 220 addresses the latency introduced by the transmission on the physical medium 46, such that the hub begins to receive the transmitted frame at a time 224 approximately coincident with the rising edge of the clock signal 214. Has a value.

The delay 220 is, for example, between the receiving circuits 78a and 78b and the nodes 42a and 42b, and between the transmitting circuits 228a and 228b and the node 42.
It can be introduced by inserting delay circuits 226a and 226b between a and 42b. See Figure 13. Since the latency of the physical medium differs from node to node depending on the length of the link, the length of the delay inserted by the circuit 226a is also appropriately changed. The calculated optimum delay can be appropriately programmed into the delay circuits 228a and 228b. This delay feature is more fully described in Applicant's U.S. patent application Ser.No. 970313, entitled "Isochronous Link Protocol," filed on the same date and incorporated herein by reference. It has been described.

The timing scheme described above ensures that the cycle received from a node arrives slightly earlier than the next cycle transmitted from the hub. A small FIFO can be inserted in the data stream received by the hub to accurately align the arrival of cycles.
A similar FIFO structure can also be used at the node to synchronize the cycle reference on which the data is received with it until it is transferred. Regarding the provision of these FIFOs, the present application entitled "Apparatus and Method for Adapting Cable Length Delay Using Isochronous FIFO", filed on the same date and incorporated herein by reference. It is described in more detail in US patent application No. 969917.

In the overall system described above, data transfer between nodes involves relaying the data from the source node to the hub, placing the data on the TSI ring interconnecting the hubs, and then transferring from the TSI ring. This is done by recovering the data at the destination station hub and routing the data from the destination station hub to the destination station node. This process wastes TSI bandwidth and introduces delay into the system if both the source and destination station nodes are connected to the same hub.

According to the invention, the circuit 58 is also provided with local loopback capability. Local loopback capability allows the circuit 58 to transmit data from the receive buffer 132 directly to the transmit buffer 154 without having to place the data on the TSI ring 134 and thus the TS.
Frees Ibus bandwidth for use by other hubs (FIGS. 2 and 3). Multiplexer 156 can be used to control loopback.

Local loopback performance is useful, for example, when both the isochronous source and the isochronous sink are connected to the same hub. For example, in the embodiment described above, both the isochronous source 48d, which is a video camera, and the isochronous sink 48b, which is a video monitor, are connected to the same hub 44a. The local loopback brings the data to the monitor 48b and sends the data received by the video camera 48d to the TSI in substantially "real time".
Display without placing on bus 134. Non-isochronous loopback can be achieved with a non-isochronous repeater device 60, such as the DP83950 Repeater Interface Controller (RIC) available from National Semiconductor, Inc. of Santa Clara, California.

FIG. 14 shows a configuration for implementing isochronous loopback. Data line 494 is normally T
It is coupled to receive transmitted data from the buffer 132 located on the SI ring 134. Data line 494
Is input to the multiplexer 156 for internal loopback, and the data on the TSI ring 134 is buffered in the buffer 154.
You are blocked from entering. Multiplexer 156 is controlled via line 160 by one of the switch tables of processor 138. The control data is stored in the output table 162 according to the destination information supplied to the signaling processor via the D channel. However, in this case, the table 162 must be wide enough to store the control bits for controlling both the TSI and the loopback multiplexer 156. When the internal loopback is initiated, the data from buffer 132a is transmitted to buffer 154 instead of timeslots. Thus, if an internal loopback is initiated in a time slot, the associated TSI data will not be loaded into buffer 154a. Multiplexer 156 is used with control line 158 indicating an internal loopback. Control of control line 158 is provided by the signaling processor and, in one embodiment, is a look-up table.

Two buffers 154a, 154b store the recovery data and are controlled in the manner previously described. During the first time frame, buffer 154a is on line 494 or TSI bus 1
Buffer 154b receives the data stored during the previous time frame from the physical medium 46b.
To line 164 of circuit 58 for transmission to the destination station node 42a via. During the next time frame, the roles of these buffers are reversed, the data received from line 494 is stored in buffer 156b, and the data stored in buffer 154a during the previous time frame is stored in the destination station node.
Output for transmission to 42a.

[0049]

As described above, according to the present invention, the use of the local loopback makes it possible to transmit data directly from the reception buffer to the transmission buffer without the need to put the data on the TSI ring. Thus T
The bandwidth of the SI bus is released for use by other hubs and the delay is reduced.

Another advantage of the loopback of the present invention is that the TSI ring 134 could potentially be omitted if a stand alone hub is used. If the data communication network consists of only a single hub, then only data line 494 is needed to connect the receive and transmit buffers of the hub. Multiplexer 156 operates in the manner described above and places the data from buffer 132 into buffer 154. Therefore, a connection to a separate TSI bus is no longer needed, but it can be provided selectively.

Although the present invention has been described in terms of a preferred embodiment, as well as specific design changes and modifications, other design changes and modifications can be used and the invention is defined by the claims. Is.

[0052]

[Table 1]

[0053]

[Table 2]

[Brief description of drawings]

FIG. 1A is a timing diagram for isochronous data frame transfer, FIG. 1B is a timing diagram for packetized data transfer,
And (C) are timing diagrams of token ring data transfer.

FIG. 2 is a block diagram of a star and ring communication system according to one embodiment of the present invention.

FIG. 3 is a block diagram of a star and ring communication system having multiple isochronous circuits in a single hub.

FIG. 4 is a block diagram of a tree-type communication system according to one embodiment of the present invention.

FIG. 5 is a block diagram of a communication system configured in accordance with one embodiment of the present invention.

FIG. 6 is a block diagram of a node circuit according to one embodiment of the present invention.

FIG. 7 is a block diagram of a hub repeater circuit according to one embodiment of the present invention.

FIG. 8 is a block diagram of a non-isochronous data reception interface according to one embodiment of the invention.

FIG. 9 is a block diagram of an isochronous data reception interface and associated hub circuitry according to one embodiment of the invention.

FIG. 10 is a block diagram of a hub transmission interface for non-isochronous data according to one embodiment of the invention.

FIG. 11 is a block diagram of a hub transmitter interface for non-isochronous data according to one embodiment of the invention.

FIG. 12 is a timing diagram for adjusting data transfer according to one embodiment of the present invention.

FIG. 13 is a block diagram of a node having a delay circuit according to one embodiment of the present invention.

FIG. 14 is a block diagram of a receive interface for isochronous data with local loopback capability according to one embodiment of the invention.

[Explanation of symbols]

 42a, 42b, 42c Node 44a Hub 46a-f Physical medium 48a, 48d Isochronous source 48b, 48e Isochronous sink 48c Non-isochronous source 48g Non-isochronous sink 48f D-channel source 50a, 50b, 50c Node circuit 54a, 54b, 54c Hub component 56 Hub circuit 494 Data line MUX multiplexer DEMUX Demultiplexer Tx transmit Rx receive isoPhy isochronous physical layer iso isochronous

 ─────────────────────────────────────────────────── ─── Continued Front Page (72) Inventor Michael Es Evans, California 95133 San Jose, Cape Hilda Place 1966 (72) Inventor Debra Jay Worsley, USA 94086 Sunnyvale, California, USA East Red Oak Drive 224-Gee

Claims (23)

[Claims] 1. A device for communicating data between a plurality of sources and sinks, the first device having a node transmitter and a node receiver coupled to at least one of the plurality of sources and sinks. A node, coupled to the node; (a) a hub receiver coupled to the node transmitter to receive data transmitted from the first node; (b) coupled to the node receiver and A hub transmitter for transmitting data to the first node; (c) a data link coupling the hub receiver to the hub transmitter; and (d) a hub between the hub receiver and the hub transmitter. A first switch disposed on the data link and coupled to the control signal, the switch having a switch for passing data from the hub receiver to the hub transmitter when the control signal is asserted;
A hub, a second hub, at least one of the hub receiver and the hub transmitter of the first hub, and the second hub, and the first hub and the second hub. A device consisting of a bus that passes data between hubs. 2. The apparatus of claim 1, wherein the first node comprises two nodes. 3. The apparatus of claim 1, wherein the switch comprises a multiplexer. 4. The apparatus of claim 1, wherein the hub receiver further comprises a buffer for storing data received from the first node. 5. The apparatus of claim 1, wherein the hub transmitter further includes a buffer that stores data for transmission to the first node. 6. The apparatus of claim 1, wherein the first hub further comprises a memory device for storing the control signal. 7. An apparatus for communicating data between a plurality of data sources and sinks, wherein at least a first one of the sources and sinks is configured to receive and transmit data isochronously. A second source and sink configured to transmit data non-isochronously, coupled to both the first and second source and sink, and (a) node transmission (B) a node receiver, and (c) is coupled to the node transmitter via a first data link for transmitting data from both the first and second ones of the source and sink, A first node having a multiplexer for providing a first dedicated bandwidth for data originating from an isochronous source, coupled to the first node, and (a) coupled to the node transmitter A hub receiver for receiving data transmitted from the first node via the first data link, and (b) transmitting data to the first node coupled to the node receiver. A hub transmitter, and (c) a second coupling the hub receiver to the hub transmitter
(D) the second link between the hub receiver and the hub transmitter.
A first hub having a switch located on the data link, coupled to the control signal, and passing data from the hub receiver to the hub transmitter when the control signal is asserted. apparatus. 8. A second hub coupled to a second node, at least one of said hub receiver and said hub transmitter of said first hub, and said second hub, 8. The apparatus of claim 7, further comprising a bus that passes data between a first hub and the second hub. 9. The apparatus of claim 7, wherein the first node comprises two nodes. 10. The apparatus of claim 7, wherein the switch comprises a multiplexer. 11. The hub receiver further includes a buffer for storing data received from the first node.
The device of claim 7. 12. The apparatus of claim 7, wherein the hub transmitter further includes a buffer that stores data for transmission to the first node. 13. The apparatus of claim 7, wherein the first hub further comprises a memory device for storing the control signal. 14. The apparatus of claim 7, wherein the second data link carries isochronous data. 15. A method of communicating data, comprising transmitting the data from a node to a hub receiver of a first hub, the data being transmitted to the hub receiver when a first control signal is asserted. From the hub to the hub transmitter via the data link, passing the data from the hub receiver to the bus, and when the second control signal is asserted, the data is routed to the second via the bus. A communication method comprising communicating to the bus. 16. The communication method of claim 15, further comprising transmitting the data from the hub transmitter to one of a plurality of nodes coupled to the first hub. 17. The method of claim 15, further comprising transmitting the data from the hub transmitter to the node. 18. A method of communicating data comprising: receiving at a first node a first set of isochronous data and a second set of non-isochronous data, the first set and the Multiplexing the second set of data to form a multiplexed data set, transmitting the multiplexed data set to a first hub via a first data link, and transmitting the multiplexed data set to the first hub. Demultiplexing at the hub receiver to obtain the isochronous data portion and the non-isochronous data portion, the isochronous data portion from the hub receiver to the hub when the first control signal is asserted. Passing through the second data link to the transmitter, passing the isochronous data portion from the hub receiver to the bus, the isochronous data portion if a second control signal is asserted. A communication from the first hub to a second hub over the bus. Communication method consisting of. 19. The communication method of claim 18, further comprising transmitting the isochronous data portion from the hub transmitter to one of a plurality of nodes coupled to the first hub. 20. The method further comprising repeating the non-isochronous data portion with a repeater device.
18 communication methods. 21. In a communication system having a plurality of hubs coupled to exchange data between a plurality of nodes, a receiving circuit for receiving isochronous data from at least one of the nodes; A transmission circuit transmitting to at least one of the nodes, a first data link coupling the reception circuit to the transmission circuit, arranged on the first data link and coupled to a control signal, A switch for passing the isochronous data from the receiver circuit to the transmitter circuit when the control signal is asserted; one of the receiver circuit or the transmitter circuit; and a given one of the plurality of hubs. A hub device coupled to the second data link for exchanging data with the given one of the plurality of hubs when a second control signal is asserted. 22. Means coupled to said receiver circuit for receiving non-isochronous data from at least one of said nodes; and non-isochronous data coupled to said at least one transmitter circuit to said at least one node. 22. The hub device of claim 21, further comprising means for transmitting the. 23. The hub circuit of claim 22, further comprising a repeater circuit coupled to the receiver circuit and transmitting the non-isochronous data from the receiver circuit to the transmitter circuit.